Whether the material falls in chunks or something closer to a liquid is highly material dependent. Dry concrete mix, coal, and soy meal (with high moisture content) certainly can, others, sand, most grains, for instance, flow and act almost like liquid.
If you knew what the recoil time of the lower hopper was, you could break up the Roark into a series of events:
If you take your Roark eq factor, which is somewhat constant (although the height of fall will decrease as the bin fills), and have = x, then the force of a 3000 lbs. chuck falling would be 3000x. If you then broke this up into a series of events, each the length of the recoil time, with the total time of the series of events equally 5 seconds, you have an analysis that comes closer to describing reality.
The recoil time, however (a factor associated with the damping of the bin) will change as the bin fills. We see this all the time with the application of rotary electric vibrators. If operated on a full bin, such a vibrator will draw 70-80% of its FLA, but if the bin is empty, then the amp draw will double to triple, sometimes worse. (Without good O/L protection, the vibrator burns up.)
Also, the effective impact area of the material on the bin will expand as the material accumulates. You can estimate how much if you have the "angle of repose" data of the material, which is the angle the material mound will form as it accumulates. That is who many of the participants rightly believe this is not a big deal; there would be many more broken hoppers if their observations did not apply.
Thus, the hopper is likely subjected to something like impacting in the early part of the 5 second pour.
No sub for testing thou, to see what model appears to describe the real world.
BK